Artificial Neural Network-Assisted Facial Analysis for Planning of Orthognathic Surgery

Area of Science:

• Oral and Maxillofacial Surgery
• Computer Science
• Medical Imaging

Background:

• Facial analysis is crucial for orthognathic surgery to achieve facial balance and correct occlusion.
• Existing methods for facial landmark detection using artificial neural networks lack specific application in orthognathic surgery planning.
• Automated facial proportion measurement is needed for precise surgical planning.

Purpose of the Study:

• To develop a computer program for automated facial measurements using photograph capture.
• To present the development stages and test results of this novel software.
• To evaluate the effectiveness of the program in orthognathic surgery planning.

Main Methods:

• Development of a computer program by researchers from the University of Pernambuco (UPE) and the Federal University of Pernambuco (UFPE).
• The program automates facial measurements through photographic capture.
• Testing and validation of the program's accuracy in detecting facial proportions.

Main Results:

• The program successfully detected nine facial measurements, including facial thirds, intercanthal distance, alar base, lip dimensions, and interlabial gap.
• Seven of the nine measurements showed high accuracy.
• Significant discrepancies (p<0.01) were noted for the middle third of the face and lower lip vermilion measurements.

Conclusions:

• The developed program is effective for automatically detecting seven facial measurements crucial for orthognathic surgery.
• Further research is needed to expand the database and enhance the artificial intelligence model for detecting more landmarks.
• Continued development will improve automated acquisition of measurements for comprehensive orthognathic surgery planning.